Vol.13 (2023) No.

1
ISSN: 2088-5334

Recognition of Agricultural Land-Use Change with Machine

Learning-Based for Regional Food Security Assessment
in Kulon Progo Plains Area
Zulfa Khoirun Nisa a, Ansita Gupitakingkin Pradipta a,*, Liana Ni’mathus Sholikah a, Bangkit Fatwa
Pratama a, Akram Sripandam Prihanantya b, Ngadisih a, Sahid Susanto a, Sigit Supadmo Arif a
a
Department of Agricultural and Biosystems Engineering, Universitas Gadjah Mada, Sleman, 55281, Indonesia
b
Department of Geodetic Engineering, Universitas Gadjah Mada, Grafika Street No. 2, Sleman, 55281, Indonesia
Corresponding author: *ansita.pradipta@ugm.ac.id

Abstract— High conversion of agricultural land in Kulon Progo Regency, as such the construction of the Yogyakarta International
Airport (YIA) and the Bedah Menoreh road, has resulted in food production and impacted food security, including Kulon Progo plains
area. This study aimed to calculate the conversion rate of agricultural land and analyze its impact on food security in the Kulon Progo
plains area from 2005 to 2020. The primary materials needed are Kulon Progo administrative maps, Landsat 7 and 8 images, land
productivity data, population data, and consumption per capita data. With tools used is Google Earth Engine (GEE), SPSS 25, Google
Earth Pro, and ArcGIS 10.3. The method used is calculating the Normalized Difference Vegetation Index (NDVI) and machine learning-
based classification through GEE to identify land-use change and analyze the state of food security. The study proved that between
2015 and 2020, there was a conversion of paddy fields, with an average rate of 126 ha/year. The existence of new paddy fields influenced
this land increase. However, in 2020 there is still food insecurity in Pengasih District, thus caused by the new paddy fields not being
optimally used for rice growth. The productivity of the land produced is not optimal. With the availability of agricultural land in 2020
(1382.85 ha), food self-sufficiency will be limited for the next 24.75 years if there is no effort to increase paddy fields.

Keywords— Land-use change; agricultural; machine learning; GEE; food security.

Manuscript received 27 Oct. 2021; revised 20 Jul. 2022; accepted 18 Sep. 2022. Date of publication 18 Feb. 2023.
IJASEIT is licensed under a Creative Commons Attribution-Share Alike 4.0 International License.

Yogyakarta International Airport (YIA) [5] and the Bedah

I. INTRODUCTION Menoreh project, the road that connects YIA and Borobudur
The use of land as a strategic and necessary asset to the Temple [6]. Agricultural land is most vulnerable to land-use
economy and to keep a sustainable living, especially amid change [7]. Agricultural land is considered very potential for
increasing food needs becoming increasingly vital [1]. As non-agricultural sector land conversion. Thus, caused by it is
time goes by, the demand for land rises; due to the increasing relatively more expansive than the other sector's land [8]. The
human population, there is an increase in food demand and high conversion rate of agricultural land to non-agricultural
consumption by the whole population [2]. The rise of the land can result in food supply and affect food security,
population makes the development of civilization and the especially rice [9]. Food security can be achieved if there is a
demands of human needs increase, accelerating regional surplus of rice supplies compared to the rice needs in a region,
development [3]. Thus, it requires additional land for various which indicates the fulfillment of food needs for everyone in
purposes such as settlements, industry, and various facilities the area [10].
and infrastructure. Despite its socio-economic significance, For minimization, it is necessary to plan land use optimally
this has led to ecological depletion and ecosystem disruptions so that sustainable land use can be carried out. Geospatial
in ecologically critical areas [4]. technology can make it easier to plan suitable land uses.
Land-use change occurs in various regions in Indonesia, Machine learning-based geospatial technology that works by
including in Kulon Progo Regency. Massive developments conducting training and learning processes or training based
currently taking place in Kulon Progo are the construction of on data collections to process data automatically has been
developed. One machine learning-based geospatial

54
technology is the Google Earth Engine (GEE) [11]. This study November or December [13]. Kulon Progo plains area is
aims to calculate the conversion rate of agricultural land and dominated by grumosol soil types with regosol in some parts
analyze its impact on food security in the Kulon Progo plains of Nanggulan and alluvial in Pengasih and Nanggulan [14].
area, which consists of the Nanggulan, Sentolo, Pengasih, and Grumusol soils have base saturation, high absorption capacity,
Lendah Districts, from 2005 to 2020. slow permeability, and erosion sensitivity [15]. A sand
fraction dominates regosol. Therefore, it has good aeration
II. MATERIALS AND METHOD and drainage. However, it has low plant available water, lousy
soil chemical properties, and significantly fewer plant
A. Study Area nutrients because they are accessible to leach [16]. Alluvial
Kulon Progo is one of five regencies located in the Special soils derived from sediment deposited in fluvial systems are
Region of Yogyakarta. Kulon Progo has topographic usually fertile and underlie the most productive agricultural
conditions between 0-1000 meters above sea level (masl), regions [17].
divided into three regions. The northern part is the highland
area and the Menoreh Hills. In the middle is a plains area with B. Data Collection
a topography classified as choppy and wavy, a transition The framework of this study can be seen in Fig. 2. The
between lowlands and hills. The southern part of Kulon Progo primary analysis is land-use change and food security,
is a lowland, a coastal area with a coast of 24.9 km. This beginning with data collection. Land-use change analysis was
research was conducted in the Kulon Progo plains area, carried out using Landsat 7 for 2005 and 2010 and Landsat 8
consisting of Nanggulan, Sentolo, Pengasih, and Lendah for 2015 and 2020. From the results of this land-use change
Districts, as shown in Fig. 1. analysis, land area is differentiated based on each land cover
class. The occurrence of this conversion of agricultural land
use will impact food production, which affects food security.
So analysis is carried out to determine the impact of changes
in agricultural land use on food security in the Kulon Progo
plains area. Data collection is required to carry out this
analysis of the initial stages of research, as shown in Table 1.
TABLE I
DATA COLLECTION
Data Year Source
Kulon Progo Department of Public Works,
administrative Housing and Settlement Areas
boundary Kulon Progo

Central Java http://tanahair.indonesia.go.id.

administrative
boundary
Landsat 7 image 2005, 2010 Google Earth Engine
Landsat 8 image 2015, 2020 Google Earth Engine
Paddy productivity 2005, 2010, Department of Agriculture and
2015 Food Kulon Progo
Total population 2005, 2010, Central Bureau of Statistics
Fig. 1 Administrative map of Kulon Progo plain area 2015, 2021 (BPS)
Kulon Progo plain area has C3 climate conditions Paddy price 2005, 2010, Central Bureau of Statistics
according to the Oldeman classification [12]. This is adequate 2015, 2020 (BPS)
to plant crops three times a year, with a couple of periods for Consumption per 2005, 2010, Central Bureau of Statistics
palawija crops on the first and second planting and paddy in capita 2015, 2020 (BPS)

Fig. 2 Research framework

55
 used as animal feed, the rice that is scattered during the
C. Analysis of Agricultural Land-Use Change milling process, and the rice as seed for subsequent planting
Land-use change is processed on the Google Earth Engine [25]. Based on [24], the determination of rice availability is
(GEE) platform, where data processing and validation described in Equation 2.
suitability tests will be carried out at this stage.
=( 1−( + + (2)
1) Calling Up Landsat Image Data: Calling Landsat 7 and
8 aim to select the image data needed by selecting the best The amount of rice expenditure per capita per month can
data for one year with the lowest cloud cover level. determine the need for rice. The value of rice consumption
expenditure per capita is then converted to the nominal price
2) Composite Image Data: Composites are carried out to of rice that applies every year to the rice consumption per
determine each year's land cover condition. There are four capita of the population. The consumption per capita will be
composites to display: the original color, water bodies, paddy multiplied by the total population to know the rice needed.
fields, and settlements. Composite image data can be obtained This value of rice needs describes the amount of rice that must
by defining the bands on each RGB channel that will be used be met to support regional food security [24].
and explaining the min and max visualization parameters [18]. The rice needs of all people in the region and the
3) Land Classification: Land classification is done by availability of rice can determine regional food security.
making training points in the classification that has been made; According to [24], the value of food security in an area can be
water bodies (blue), vegetation (dark green), paddy fields formulated by Equation 3.
(light green), built-up land (red), and open land (yellow). This Food security = Availability of rice − Need for rice (3)
training point is used as an algorithm defined in the machine
learning process. Food security with a surplus value in an area indicates a
higher level of regional food security. The higher the food
4) Normalized Difference Vegetation Index: NDVI is a deficit that occurs, the lower the level of food security. If the
calculation analysis method used to identify vegetation result shows a positive number (surplus), it indicates food
density by observing the color displayed on the image. security, and vice versa; if it shows a negative value (deficit),
Calculations on NDVI using band 4 (Red) and band 5 (Near- it indicates food insecurity [26].
infrared) are reflected by the leaves [19]. The value of the
NDVI calculation can be read as the closer the value to -1, the E. Analysis of Self-sufficiency Limit
lower the vegetation level, then the closer the value to +1, the Rice self-sufficiency is an effort to fulfill local and national
higher the vegetation level. And if the resulting value is close food needs. The limit of self-sufficiency in rice can be
to or less than 0, there is no vegetation in the area [20]. interpreted as a limit of an area that can still meet the rice
5) Classification Suitability Test: The classification needs of the population in its territory [27]. The analysis of
suitability test attempts to determine the accuracy of the impact of land use on rice self-sufficiency can be
classification results that have been carried out. This test aims calculated as Equation 4.
to approximate the classification results to the actual data and 3 4 56 4 78 4 9
= :
(4)
determine the confidence level in using classification results
for analysis and further purposes [21]. There are several ways The predicted value of paddy fields will be obtained from
to carry out a classification suitability test, particularly field the equation above. The paddy field area number will be used
validation and confusion matrix tests. The confusion matrix to calculate the rice self-sufficiency limit. The rate of land-use
test calculates the presence or absence of errors in each form change with the population growth rate can be described as
of land cover from the results of the image classification mutually influencing each other to form the rice self-
process [22]. A confusion matrix is a square matrix containing sufficiency limit, as described in Equation 5 [27].
the number of pixels classified in overall accuracy, kappa
( ;4
accuracy, producer accuracy, and user accuracy. The land = 1 (5)
cover classification accuracy is more than 80% [23]. Then, if it is known the value of the rate of population
D. Analysis of Food Security growth against the requirement for agricultural land, it will
produce an equation due to the population growth rate in an
Regional food security is achieved when there is a surplus exponential function such as Equation 6. This is because the
of rice supply than the need for rice, which indicates food population growth rate is an exponential function.
needs for everyone in the region [10]. Food security begins
( ;4
with calculating the area in each district in the Kulon Progo = 2 (6)
plains area obtained from the analysis of land changes. Paddy
One cut point (x, y) will be generated as a result of applying
production can be calculated based on [24] by multiplying the
Equations 5 and 6, and this point can be interpreted as the limit
land area and the existing paddy productivity as described in
of rice's self-sufficiency. Imagine that there is a consistent
Equation 1.
shift in land usage while at the same time the pace of
= (1) population growth contributes to an ever-increasing need for
agricultural land. In that scenario, there will be a situation
Later, we determined the calculation of the availability of where there is no longer sufficient food security.
rice. The measure of rice availability aims to convert changes
from paddy to rice. This conversion is based on the value of
paddy shrinkage when it is processed into rice because it is

56
F. Multiple Linear Statistical Test F-test is carried out to determine the effect of the
The indicators that this approach of ensuring food security independent variables simultaneously on the dependent
is acceptable can be observed from the regional food security variable. If the probability of the value of F significance <0.05,
value, which is obtained by searching for the difference it can be said that there is an influence between the
between the amount of food that is available and the amount independent variables on the dependent variable
of food that is required. There are three classifications of food simultaneously and the other way around. A T-test was
security; food security occurs when the difference in rice carried out to test each variable partially. If the significance of
availability and rice demand shows a positive value (+), and the T value < 0.05, it can be said that there is an influence
sufficient food occurs when the difference shows zero (0). between the independent variables on the dependent variable
Food insecurity occurs when the difference shows a negative partly and the other way around.
value (-) [26].
Then multiple linear statistical tests were conducted to III. RESULTS AND DISCUSSION
determine the relationship between land conversion and food A. Agricultural Land-Use Change
security. This is performed to obtain the coefficient of
determination or R. The R-value is the independent variable 1) Normalized Difference Vegetation Index: The
(x) effect on the dependent variable (y). The independent vegetation level changes every year. The denser the level of
variable (x) was tested in paddy fields, and population growth green in an area indicates, the thicker the vegetation in the
conversion, and the related variable (y) was in rice production. area.
In addition, the coefficient of determination can be used to
2) Land Classification: The classification process is carried
predict and check how much influence the variable x
out using GEE with the CART method, a decision tree-based
simultaneously contributes to the y variable [28]. Based on
algorithm [29]. The classification of land cover classes can be
scostatistical testing can be done using SPSS (Statistical
seen in Fig. 3.
Package for Social Science), consisting of several stages such
as the F test and T-test.

The year 2005 The year 2010

The year 2015 The year 2020

Fig. 3 Visualization of land cover class

57
 Paddy production can be calculated by multiplying the
3) Classification Suitability Test: The confusion matrix test
existing land area and paddy productivity as described in
calculates the errors of land cover from image classification
Equation 1, which obtains paddy production data, as shown in
results. The confusion matrix is a square matrix containing a
Table 2.
number of pixels classified in overall accuracy, kappa
accuracy, producer accuracy, and user accuracy. Allowed TABLE II
CALCULATION FOOD NEEDS
classification accuracy is more than 80% [30]. Six different
spots are selected at random and used to create a confusion D p R A N
matrix. The findings indicate that the kappa accuracy and Year 2005
ground checking in 2005, 2010, 2015, and 2020 are exhibiting Lendah 51469.41 30154.07 29149.94 28571.48
more than 80 percent per year from the two methods used to Sentolo 66003.39 38669.01 37381.33 35105.99
Pengasih 71758.13 42040.5 40640.55 36757.17
test the accuracy: the Google Earth Engine and ground check.
Nanggulan 52931.65 31010.75 29978.09 24722.20
According to this result, the accuracy acquired is pretty Year 2010
accurate, with the land cover classification findings mirroring Lendah 34899.29 20284.48 19609.01 26837.74
the scenario observed in the field. Sentolo 46608.64 27090.29 26188.18 32785.98
Pengasih 52245.52 30366.61 29355.4 33264.60
B. Land Conversion Rate Nanggulan 67275.18 39102.289 37800.18 20057.43
Land cover in the Kulon Progo plains area can be divided Year 2015
Lendah 57455.85 33395.01 32282.95 28590.56
into five classes: water bodies, vegetation, paddy fields, built- Sentolo 85539.26 49717.89 48062.29 35111.12
up land, and open land. After classifying each land cover, the Pengasih 51090.97 29695.55 28706.69 35693.85
area of each land cover class can be known. The analysis Nanggulan 77036.59 44775.9 43284.87 21379.64
obtained shows that the water body cover area tends to Year 2020
increase every year, except in 2020, which decreases. In 2015 Lendah 75377.74 43839.68 42379.81 27811.18
Sentolo 133376.45 77571.72 74988.58 34430.42
the increase was influenced by the construction of the Bogor Pengasih 59034.37 34334.38 33191.05 36189.81
reservoir. Then the decrease in 2020 is caused by Nanggulan 91108.69 52988.79 51224.27 20921.09
sedimentation or siltation due to waste disposal. The land
cover change of vegetation, paddy field, built-up land, and Paddy production data is then converted into rice
open land was significant from 2005 to 2010, which increased production using Equation 2. It considers the correction factor
and then decreased in 2015. Holes in the Landsat 7 image for paddy loss obtained from BPS Kulon Progo data, such as
might cause this fluctuation in the area because the Scan Line paddy to be used as seed, paddy for animal feed, scattered
Corrector (SLC), the instrument to compensate for forwarding paddy, and paddy for non-industrial food. Then, food
satellite movement, encountered a permanent failure. So then, availability is calculated by considering paddy production
the resulting pattern is zig-zag, which causes gaps to be with correction factors for rice loss. The availability of food
generated in the image data [31]. It causes the resulting in quintal units is obtained from this calculation in the Kulon
interpretation to be less accurate, and there is still a lot of Progo plains area.
cloud cover that confuses image reading. Also, a new paddy The valuation of each district's food requirements can be
fields printing program in 2015 influenced the increase in the calculated by multiplying the amount of rice consumed by that
paddy fields area. district's entire population. The demand for rice is initially
expressed in kilograms per year, which is subsequently
C. Food Security Condition converted to quintals per year. The regional food security
Data on the productivity of paddy fields in the Kulon Progo value is determined by looking for the difference between
plains area in 2005, 2010, and 2015 were obtained from the food availability and food needs. There are three
Central Bureau of Statistics (BPS) of Kulon Progo. Data for classifications of food security; food security occurs when the
the productivity of paddy fields in 2020 were obtained from difference in rice availability and rice demand shows a
linear extrapolation calculations (shown in Fig. 4). positive value (+), and sufficient food occurs when the
difference shows zero (0). Food insecurity occurs when the
90
difference shows a negative value (-) [26]. The food security
80
Rice productivity (quintal/ha)

70
condition in the Kulon Progo plains is shown in Fig. 5.
60 All districts in the Kulon Progo plains area experienced
50 food security conditions in the year 2005, and this was the
40 case throughout all of the region's districts. In 2010, the
30 Nanggulan District was in a position of food security, while
20 the Pengasih, Lendah, and Sentolo Districts were in a position
10 of food insecurity. A lack of completeness in the image data
0 may be to account for the change from food security in 2005
Lendah Sentolo Pengasih Nanggulan to food insecurity in 2010; an inaccurate interpretation of the
Districts data could have caused this. Nanggulan, Sentolo, and Lendah
Districts experienced food security circumstances in 2015 and
2005 2010 2015 2020 2020; while Pengasih Districts suffered conditions of food
Fig. 4 Land productivity
insecurity during the same time period.

58
 Fig. 5 Food security map

The newly built rice field can cause this food insecurity an accurate prediction of this self-sufficiency limit. Because
condition. It causes the new paddy fields not to be optimally the plains area of Kulon Progo does not export and does not
used for rice growth so that the productivity of the new paddy get imports from other regions, it is assumed that the
land is not maximized. The new paddy fields also need good population is stable; it does not fluctuate, and the rate of land
irrigation support to meet their water needs. Meanwhile, conversion is expected to always decline. Neither of these
according to the Department of Agriculture and Food Kulon factors contribute to population growth.
Progo some of the new printed paddy fields in the Pengasih
TABLE III
District do not have adequate irrigation. According to [32], PREDICTION OF PADDY FIELD CONDITIONS
stabilizing the new paddy fields ecosystem takes
approximately ten years. So that although the area of paddy District Paddy Field Paddy Field Informa
2020 (ha) Needed (ha) tion
fields increases with the presence of new paddy fields, the
Lendah 990.709 291.305 Surplus
quality of the new paddy fields tends to decrease. In addition, Sentolo 1713.821 352.577 Surplus
it can also be caused by external factors such as climate Pengasih 810.387 395.912 Surplus
change and disturbances from pests. Nanggulan 1120.999 205.142 Surplus
D. Land Transfer Function Impact on Food Security Based on Table 3, Lendah, Sentolo, Pengasih, and
Multiple linear statistical tests were conducted to determine Nanggulan Districts have more paddy fields than the required
the land conversion impact on food security. Regression paddy fields. Furthermore, the calculation of the self-
analysis is a data analysis technique in statistics used to sufficiency limit is carried out. The self-sufficiency limit can
examine the relationship between several variables [33]. The be known by looking at the intersection between the graph of
f-test was conducted to determine the effect of the the rate of land conversion and the graph of population growth
independent variables simultaneously on the dependent rate. This cut-off point describes a condition of the inability
variable. The F test results obtained a significance of more of an area to meet the rice needs in that area. Prediction of
than 0.05; this indicates that the conversion of paddy fields self-sufficiency limit is made by referring to Equation 5 and
and population growth together have no significant effect on Equation 6. Wherefrom this calculation for the conversion of
rice production. Then T-test was conducted to test each agricultural land functions is ( = 4635.96( CD.DEEFGH ,
variable partially. The T-test results are more than 0.05; it where the value 4635.96 is the existing land area, and the
indicates no considerable impact between each independent value -0.04485 indicates the rate of reduction of agricultural
variable, the conversion of paddy fields, and population land. Then, the population growth rate equation to the need
growth to the dependent variable, particularly rice production.
for agricultural land is ( = 1244.9( D.DDEIEJH . Where
E. Self-Sufficiency Limit Prediction 1244.9 is the value of land needed, 0.004243 is the population
The rice self-sufficiency limit is defined as a limit for an growth rate. The limit of food self-sufficiency can be seen in
area that is still fulfilled the rice needs of the population in its Fig. 6.
area [27]. A number of assumptions are made in order to make

59
 5000 Pr paddy fields productivity kg/ha
Agricultural land conversion rate (ha)

4500 Pl rotation of rice crops in a year planting/

4000 Self-Sufficiency Limit (24.75 ; rice/year
3500 1382.85) R yield of rice or shrinkage of grain into 1/100
3000 rice
2500 K average consumption of rice per person kg/prsn/yr
2000 in a year
1500 Fx the need for paddy fields at the rate of ha
1000 conversion of agricultural land
500 a1 existing paddy fields area ha
0 fx existing paddy fields area ha
0 20 40 60 80 a2 land requirement ha
Time (Year) x time is taken years
D district
Agricultural land conversion rate (ha)
p paddy production quintal
Population growth rate (People)
R rice production quintal
A rice availability quintal
Fig. 6 Food self-sufficiency limit N food needs quintal
This cut-off point indicates that in 1382.85 ha and 24.75
years in the future, the Kulon Progo plain area can only meet ACKNOWLEDGMENT
the need for food for its residents or cannot export. This will We are entirely grateful to the Research Directorate of
occur if there is not an increase in paddy fields either through Universitas Gadjah Mada, who provide the funding system for
intensification, which refers to efforts to increase agricultural this study through the grant of final project recognition. We
yields by optimizing agricultural land to obtain optimal also thank the Department of Agriculture and Food and
results, or extensification, which refers to expanding land by Central Bureau of Statistics of Kulon Progo Regency, who
opening up new land that can be planted with crops. Both of provide data and information related to this research.
these methods are referred to as land optimization.
REFERENCES
IV. CONCLUSION [1] M. H. Jamil, R. M. Rukka, A. N. Tenriawaru, R. Achmad, A. T.
Nugraha, and Y. T. Walangadi, “The existence of rice fields in
The establishment of the Yogyakarta International Airport Makassar City,” in IOP Conference Series: Earth and Environmental
(YIA) and the Bedah Menoreh road in Kulon Progo has Science, 2021, vol. 681, no. 1, doi: 10.1088/1755-1315/681/1/012091.
resulted in the rapid conversion of agricultural land. This has [2] S. B. Khade, R. S. Khillare, and M. B. Dastagiri, “Global Livestock
Development: Policies and Vision,” Indian J. Anim. Sci., vol. 91, no.
affected food production and security, particularly in the 9, pp. 770–779, 2021.
Kulon Progo plains area. The essential materials required are [3] S. Liu, S. Gao, W. L. Hsu, Y. C. Shiau, and H. L. Liu, “Mechanism
administrative maps of Kulon Progo, Landsat 7 and 8 images, study on the impact of china population structure change on the water
and statistics on land productivity, population, and use of the three main industries,” Sustain., vol. 14, no. 1, 2022, doi:
10.3390/su14010204.
consumption per capita. Machine learning was utilized to [4] K. T. Deribew, “Spatiotemporal analysis of urban growth on forest and
classify land-use changes and look at the state of food security. agricultural land using geospatial techniques and Shannon entropy
Google Earth Engine (GEE) is the main tool used for this. method in the satellite town of Ethiopia, the western fringe of Addis
The results showed that the conversion rate of agricultural Ababa city,” Ecol. Process., vol. 9, no. 1, 2020, doi: 10.1186/s13717-
020-00248-3.
land in the Kulon Progo plains from 2005 to 2010 decreased [5] D. A. Puspitaningrum, “System dynamic modelling of agriculture land
by 97.6 ha/year; in the period 2010 to 2015, it increased by availability,” in IOP Conference Series: Earth and Environmental
108.8 ha/year; in the period 2015 to 2020, it was 126.02 ha Science, 2019, vol. 250, no. 1, p. 2022, doi: 10.1088/1755-
/year. It resulted in the food security status in 2010 in Pengasih, 1315/250/1/012087.
[6] L. N. Sholikah et al., “Identification of agricultural land use change
Lendah, and Sentolo Districts being categorized as food based on machine learning for regional food security analysis in the
insecure; in 2015 and 2020, food security conditions in mountainous region of Kulon Progo regency,” in IOP Conference
Pengasih District were found to be in the unsure food category. Series: Earth and Environmental Science, 2021, vol. 922, no. 1, p.
The productivity of the land produced is not optimal. With the 012060, doi: 10.1088/1755-1315/922/1/012060.
[7] S. A. Purba, B. Slamet, and A. Rauf, “Spatial Modelling of Land
availability of agricultural land in 2020 (1382.85 ha), food Conversion Vulnerability In Padang Watersheds North Sumatera
self-sufficiency will be limited for the next 24.75 years if there Province,” in IOP Conference Series: Earth and Environmental
is no effort to increase paddy fields. Science, 2021, vol. 782, no. 3, p. 2022, doi: 10.1088/1755-
1315/782/3/032026.
[8] A. Garud and B. Rao, “Understanding the Implications of the Loss of
NOMENCLATURE Peri-Urban Arable Land—A Case of Pune Metropolitan Region,” in
p paddy production quintal/yr Lecture Notes in Civil Engineering, 2021, vol. 121 LNCE, pp. 433–
445, doi: 10.1007/978-981-33-4114-2_35.
S seeds [9] F. Firmansyah, C. Susetyo, N. A. Pratomoatmojo, U. F. Kurniawati,
F animal feed and M. Yusuf, “Land Use Change Trend of Paddy Field and Its
W scattered Influence on Food Security In Gerbangkertosusila Region,” in IOP
C conversion of paddy to rice Conference Series: Earth and Environmental Science, 2021, vol. 778,
no. 1, doi: 10.1088/1755-1315/778/1/012023.
P total population people [10] V. K. Nguyen, D. Dumaresq, and J. Pittock, “Impacts of rice
L area of paddy fields ha intensification on rural households in the Mekong Delta: emerging

60
 relationships between agricultural production, wild food supply and 10.1007/s12524-021-01398-3.
food consumption,” Food Secur., vol. 10, no. 6, pp. 1615–1629, 2018, [23] A. Hedayati, M. H. Vahidnia, and S. Behzadi, “Paddy lands detection
doi: 10.1007/s12571-018-0848-6. using Landsat-8 satellite images and object-based classification in
[11] A. Raj and L. K. Sharma, “Assessment of land-use dynamics of the Rasht city, Iran,” Egypt. J. Remote Sens. Sp. Sci., vol. 25, no. 1, pp.
Aravalli range (India) using integrated geospatial and CART approach,” 73–84, 2022, doi: 10.1016/j.ejrs.2021.12.008.
Earth Sci. Informatics, vol. 15, no. 1, pp. 497–522, 2022, doi: [24] I. M. Y. Prasada, A. Dhamira, and Masyhuri, “The potential loss of
10.1007/s12145-021-00753-9. rice production due to wetland conversion in East Java,” in IOP
[12] H. Mustaqim, “Rainfall Analysis for Meteorological Drought in Kulon Conference Series: Earth and Environmental Science, 2019, vol. 230,
Progo Regency 2006-2015,” Universitas Muhamadiyah Surakarta, no. 1, pp. 0–6, doi: 10.1088/1755-1315/230/1/012005.
2016. [25] D. Nofriati, N. Asni, and S. Primilestari, “The Study of Paddy Harvest
[13] M. Qonita, “Agricultural planning based on local agro-climatology Losses Determination on Tidal Land in Tanjung Jabung Timur Region
assessment in Bongkot, Purwodadi, Purworejo,” in IOP Conference Jambi Province,” in IOP Conference Series: Earth and Environmental
Series: Earth and Environmental Science, 2021, vol. 686, no. 1, doi: Science, 2019, vol. 309, no. 1, pp. 0–4, doi: 10.1088/1755-
10.1088/1755-1315/686/1/012052. 1315/309/1/012017.
[14] RPI2-JM Randal Task Force Team for DIY Creative Works, “RPI2- [26] A. A. Adenle, K. Wedig, and H. Azadi, “Sustainable agriculture and
JM KULON PROGO REGENCY 2015-2019,” Yogyakarta, 2014. food security in Africa: The role of innovative technologies and
[15] S. E. Prasetyowati and Y. Sunaryo, “Effect of ameliorants on canopy international organizations,” Technol. Soc., vol. 58, no. April, p.
architectures of jack bean (Canavalia ensiformis) cultivated in 101143, 2019, doi: 10.1016/j.techsoc.2019.05.007.
marginal soils,” in IOP Conference Series: Earth and Environmental [27] R. Martanto, Analysis of Land Use Change Patterns for Rice Self-
Science, 2021, vol. 681, no. 1, doi: 10.1088/1755-1315/681/1/012036. Sufficiency Stability in Sukoharjo Regency. Yogyakarta: STPN Press,
[16] G. Gusnidar, F. Fitria, L. Maira, and Y. Yulnafatmawita, “Role of 2019.
compost derived from rice straw and tithonia in improving chemical [28] E. F. Akmam, T. Siswantining, S. M. Soemartojo, and D. Sarwinda,
fertility of Regosol on onion cultivation,” in IOP Conference Series: “Multiple Imputation with Predictive Mean Matching Method for
Earth and Environmental Science, 2019, vol. 347, no. 1, doi: Numerical Missing Data,” in ICICOS 2019 - 3rd International
10.1088/1755-1315/347/1/012095. Conference on Informatics and Computational Sciences: Accelerating
[17] L. Mao et al., “Improved geochemical baseline establishment based on Informatics and Computational Research for Smarter Society in The
diffuse sources contribution of potential toxic elements in agricultural Era of Industry 4.0, Proceedings, 2019, p. 2022, doi:
alluvial soils,” Geoderma, vol. 410, no. July 2021, p. 115669, 2022, 10.1109/ICICoS48119.2019.8982510.
doi: 10.1016/j.geoderma.2021.115669. [29] N. Ponganan, T. Horanont, K. Artlert, and P. Nuallaong, “Land Cover
[18] L. Wang, J. Wang, and F. Qin, “Feature fusion approach for temporal Classification using Google Earth Engine’s Object-oriented and
land use mapping in complex agricultural areas,” Remote Sens. Machine Learning Classifier,” in 2021 2nd International Conference
(Multidisciplinary Digit. Publ. Institute), vol. 13, no. 13, 2021, doi: on Big Data Analytics and Practices, IBDAP 2021, 2021, pp. 33–37,
10.3390/rs13132517. doi: 10.1109/IBDAP52511.2021.9552099.
[19] S. Karki et al., “Mapping Spatial Distribution and Biomass of [30] N. H. Quang et al., “Multi-decadal changes in mangrove extent, age
Intertidal Ulva Blooms Using Machine Learning and Earth and species in the Red River Estuaries of Viet Nam,” Remote Sens.
Observation,” Front. Mar. Sci., vol. 8, no. April, pp. 1–20, 2021, doi: (Multidisciplinary Digit. Publ. Institute), vol. 12, no. 14, 2020, doi:
10.3389/fmars.2021.633128. 10.3390/rs12142289.
[20] Z. E. Kulenbekov, S. Z. Orunbaev, and B. D. Asanov, “Investigation [31] N. Case and A. Vitti, “Reconstruction of multi-temporal satellite
of the High Mountain Vegetation Using Satellite Imagery, Kyrgyzstan,” imagery by coupling variational segmentation and radiometric
Springer Water, pp. 151–168, 2021, doi: 10.1007/978-3-030-68337- analysis,” ISPRS Int. J. Geo-Information, vol. 10, no. 1, 2021, doi:
5_15. 10.3390/ijgi10010017.
[21] H. Ismanto, A. Doloksaribu, D. S. Susanti, and D. F. Septarini, “The [32] B. Irawan and S. Friyanto, “The Impact of Rice Field Conversion in
accuracy of remote sensing image interprepation on changes in land Java on Rice Production and Its Control Policy,” no. 1, pp. 1–33, 2002.
use suitability in merauke regency papua,” in International Journal of [33] J. Pan, “Improved two-stage model averaging for high-dimensional
Engineering Trends and Technology, 2020, vol. 68, no. 10, pp. 42–47, linear regression, with application to Riboflavin data analysis,” BMC
doi: 10.14445/22315381/IJETT-V68I10P207. Bioinformatics, vol. 22, no. 1, pp. 1–17, 2021, doi: 10.1186/s12859-
[22] M. Bamdadinejad, M. J. Ketabdari, and S. M. H. Chavooshi, 021-04053-3.
“Shoreline Extraction Using Image Processing of Satellite Imageries,”
J. Indian Soc. Remote Sens., vol. 49, no. 10, pp. 2365–2375, 2021, doi:

61